Can I run Llama 3.3 70B on a Mac?

Yes: Llama 3.3 70B runs on 25 of 54 Macs in the LLMCheck index. It needs about 44 GB at a 4K context. Smallest current Mac: the Mac mini M5 Pro 48 GB, raised GPU limit; cheapest at 15+ tok/s: the Mac Studio M5 Ultra 96 GB ($5,499; 25 tok/s published). Best measured: 25.2 tok/s on an M5 Ultra (community).

RAM at 4K
44
GB
RAM at 64K
—
no KV figure
Macs that run it
25
of 54 configs
Best measured
community25.2
tok/s · M5 Ultra

Llama 3.3 70B on every Mac

One row per Mac configuration, grouped by product line. The build is the one each speed is for; the context column is the longest prompt that fits inside macOS’s default GPU memory limit with a 16-bit KV cache (“raised” where the limit has to be raised first; “full” when the whole window fits). Speeds are decode tok/s from the LLMCheck index, each with its provenance mark. For dense models, the formula’s typical error is about 9% (the median gap to 75 measured plain-decode runs; it tends to undershoot (median −6%), and an estimate scaled from the model’s own measured run is usually closer; method).

Runs fits macOS’s default GPU limit, 15+ tok/sRuns, slowly fits, under 15 tok/sRaise GPU limit fits once the limit is raised (command below)Doesn't fit needs more memory than the Mac has
MacVerdictBuildSpeedMax contextModel page
MacBook Air
MacBook Air M5 16 GBDoesn't fit————
MacBook Air M5 24 GBDoesn't fit————
MacBook Air M5 32 GBDoesn't fit————
MacBook Air M1 8 GB refurbDoesn't fit————
MacBook Air M4 16 GB usedDoesn't fit————
MacBook Air M2 16 GB refurbDoesn't fit————
MacBook Air M3 16 GB refurbDoesn't fit————
MacBook Air M1 16 GB refurbDoesn't fit————
MacBook Air M4 24 GB usedDoesn't fit————
MacBook Air M2 24 GB refurbDoesn't fit————
MacBook Air M3 24 GB refurbDoesn't fit————
MacBook Pro
MacBook Pro M6 16 GBDoesn't fit————
MacBook Pro M5 Pro 24 GBDoesn't fit————
MacBook Pro M6 24 GBDoesn't fit————
MacBook Pro M6 32 GBDoesn't fit————
MacBook Pro M5 Pro 48 GBRaise GPU limit command ↓Q4_K_Mestimated6 tok/s——
MacBook Pro M5 Max 64 GBRuns, slowly4-bitsourced13 tok/s—M5 Max page
MacBook Pro M5 Pro 64 GBRuns, slowlyQ4_K_Mestimated6 tok/s——
MacBook Pro M5 Max 128 GBRuns, slowly4-bitsourced13 tok/s—M5 Max page
MacBook Pro M3 Max 64 GB refurbRuns, slowlyQ4_K_Mestimated8 tok/s——
MacBook Pro M3 Max 128 GB refurbRuns, slowlyQ4_K_Mestimated8 tok/s——
Mac mini
Mac mini M6 16 GBDoesn't fit————
Mac mini M5 Pro 24 GBDoesn't fit————
Mac mini M6 24 GBDoesn't fit————
Mac mini M6 32 GBDoesn't fit————
Mac mini M5 Pro 48 GBRaise GPU limit command ↓Q4_K_Mestimated6 tok/s——
Mac mini M5 Pro 64 GBRuns, slowlyQ4_K_Mestimated6 tok/s——
Mac mini M2 Pro 16 GB refurbDoesn't fit————
Mac mini M4 16 GB usedDoesn't fit————
Mac mini M4 Pro 24 GB usedDoesn't fit————
Mac mini M4 24 GB usedDoesn't fit————
Mac mini M2 Pro 32 GB refurbDoesn't fit————
Mac mini M4 Pro 48 GB usedRaise GPU limit command ↓Q4_K_Mestimated5 tok/s——
Mac mini M4 Pro 64 GB usedRuns, slowlyQ4_K_Mestimated5 tok/s——
iMac
iMac M6 16 GBDoesn't fit————
iMac M6 24 GBDoesn't fit————
iMac M6 32 GBDoesn't fit————
Mac Studio
Mac Studio M5 Max 36 GBDoesn't fit————
Mac Studio M5 Max 48 GBRaise GPU limit command ↓4-bitsourced13 tok/s—M5 Max page
Mac Studio M5 Max 64 GBRuns, slowly4-bitsourced13 tok/s—M5 Max page
Mac Studio M5 Ultra 96 GBRuns4-bitsourced25 tok/s—M5 Ultra page
Mac Studio M5 Max 128 GBRuns, slowly4-bitsourced13 tok/s—M5 Max page
Mac Studio M5 Ultra 256 GBRuns4-bitsourced25 tok/s—M5 Ultra page
Mac Studio M5 Ultra 512 GBRuns4-bitsourced25 tok/s—M5 Ultra page
Mac Studio M2 Max 32 GB refurbDoesn't fit————
Mac Studio M1 Ultra 64 GB refurbRunsQ4_K_Mestimated15 tok/s——
Mac Studio M4 Max 64 GB usedRuns, slowlyQ4_K_Mestimated12 tok/s——
Mac Studio M2 Max 64 GB refurbRuns, slowlyQ4_K_Mestimated8 tok/s——
Mac Studio M1 Ultra 128 GB refurbRunsQ4_K_Mestimated15 tok/s——
Mac Studio M2 Ultra 128 GB refurbRunsQ4_K_Mestimated15 tok/s——
Mac Studio M4 Max 128 GB usedRuns, slowlyQ4_K_Mestimated12 tok/s——
Mac Studio M2 Ultra 192 GB refurbRunsQ4_K_Mestimated15 tok/s——
Mac Studio M3 Ultra 256 GB usedRunsQ4_K_Mestimated16 tok/s——
Mac Studio M3 Ultra 512 GB refurbRunsQ4_K_Mestimated16 tok/s——
Provenance key
sourced
Published figure — vendor or press review
estimated
community
Linked individual run — blog, forum or submission

The index has no KV-cache figure for Llama 3.3 70B yet, so the context column shows —: its minimum RAM covers a short (about 4K-token) prompt only.

Raising the GPU memory limit

macOS lets the GPU wire only part of unified memory by default (Metal’s recommended working set; M5 and M6 Macs run macOS 26 or later, whose default is higher), and Llama 3.3 70B needs about 44 GB. On the Macs marked Raise GPU limit it fits in RAM but not inside that default. The default is a performance threshold, not a hard wall: past it, the runtime moves part of it to the CPU or shortens the context (llama.cpp, Ollama), runs it slowly (MLX), or stops with an out-of-memory error. One command raises the limit until the next restart (sudo sysctl iogpu.wired_limit_mb=0 restores the default sooner); it leaves macOS a reserve, and other apps get less memory while the model is loaded.

Llama 3.3 70B: memory by quantization

Weights are 70B parameters at each format’s effective size (the index’s GB-per-billion table), plus about 2 GB for the runtime and a short context; longer contexts add the KV cache computed from the model’s config.json at 16 bits. The first row is the catalog’s own figure (about 44 GB at a 4K context), the one every verdict on this page uses. “Smallest Mac” is the smallest memory size among the Macs Apple sells new that holds it at a 4K context inside that Mac’s default GPU limit. Estimates, not measurements.

BuildWeightsRAM at 4Kat 32Kat 64KSmallest Mac
As catalogued—44 GB——64 GB
Q2_K23.8 GB26 GB——36 GB (32 GB with a raised limit)
Q3_K_M34.3 GB36 GB——48 GB
Q4_K_M39.9 GB42 GB——64 GB (48 GB with a raised limit)
Q5_K_M49.7 GB52 GB——64 GB
Q6_K57.4 GB59 GB——96 GB
Q8_074.2 GB76 GB——96 GB
F16140.0 GB142 GB——256 GB

Measured runs of Llama 3.3 70B

Every published (filled square) and community (hollow circle) run the index holds for Llama 3.3 70B: 5 published, 1 community, each linked to its source. LLMCheck runs no tests of its own. One row is one measurement session; on each chip the table above shows the lower median of the eligible runs (plain decode at 5 bits or fewer), and a binned or “not an anchor” run is shown but never used as a reference.

ChipRuntimeQuantDecodeContextDateSource
M5 Ultra 256 GBMLX4-bitcommunity25.2 tok/s512-token random prompt, 128 generated, mean of 32026-11-11bohemiansilicon.com
M5 Ultra 256 GBMLX4-bitsourced25 tok/sempty context, 128 generated2026-10-26vsesmart.cz
M5 Ultra 256 GBllama.cpp4-bitsourced23 tok/sempty context, 128 generated2026-10-26vsesmart.cz
M5 Max 128 GBMLX4-bitsourced13.6 tok/srandom-token prompt, empty context2026-11-02vsesmart.cz
M5 Max 128 GBMLX4-bitsourced13.1 tok/s—2026-09-21tbreak.com
M5 Max 128 GBllama.cpp4-bitsourced12.5 tok/srandom-token prompt, empty context2026-11-02vsesmart.cz

Recommended setup

The quickest start is Ollama. Pull the model by its tag (checked on ollama.com, 2026-11-14; the file is about 43 GB):

ollama run llama3.3:70b

Prefer a window to a terminal? LM Studio downloads the same model with one click. · model card →

On the Mac mini M5 Pro (64 GB), the smallest current Mac that runs it inside the default GPU limit, the index figure is ~6 tok/s (estimated) at Q4_K_M; keep the context at or under a short prompt there, or use a q8 KV cache. The llama3.3:70b file fits there.

Step-by-step: the Ollama install guide.

🛒 Which Mac to get for Llama 3.3 70B

Mac mini M5 Pro (48 GB) — smallest current Mac that runs it, $2,299 new (Apple list price, checked 2026-11-14), ~6 tok/s (estimated) once the GPU memory limit is raised. For Ollama: the llama3.3:70b tag is about 43 GB, so it needs a 64 GB Mac; on 48 GB use a ~4-bit build after raising the GPU limit. Mac Studio M5 Ultra (96 GB) — cheapest at 15+ tok/s, $5,499 new (Apple list price, checked 2026-11-14), 25 tok/s (published figure). Amazon does not stock the Mac mini M5 Pro (48 GB): Apple builds it to order. Amazon does not stock the Mac Studio M5 Ultra (96 GB): Apple builds it to order.

FAQ: Llama 3.3 70B on a Mac

Can I run Llama 3.3 70B on a Mac?

Yes. Llama 3.3 70B needs about 44 GB of unified memory at a short (4K-token) context, so it runs on 25 of the 54 Mac configurations in the LLMCheck index; 21 of them hold it inside macOS’s default GPU memory limit and 4 only after raising that limit. The smallest Mac Apple sells new that runs it is the Mac mini M5 Pro (48 GB).

How much RAM does Llama 3.3 70B need?

About 44 GB at a short (about 4K-token) context: the weights plus about 2 GB for the runtime. The index has no KV-cache figure for Llama 3.3 70B yet, so it cannot say how much a long context adds.

How fast does Llama 3.3 70B run on a Mac?

On an M5 Max, the chip the leaderboard quotes it on, the index figure is 13 tok/s (published figure) at 4-bit. The best measured figure is 25.2 tok/s on an M5 Ultra 256 GB (community run, 4-bit). Every Mac’s figure is in the table on this page with its provenance: published, community run or estimated.

What is the cheapest Mac that runs Llama 3.3 70B well?

Among Macs Apple sells new, the Mac Studio M5 Ultra (96 GB), $5,499 new (Apple list price, checked 2026-11-14), is the cheapest that reaches 15 tok/s with Llama 3.3 70B, at 25 tok/s (published figure).

Do I need to raise the GPU memory limit to run Llama 3.3 70B?

On 4 of the 54 configurations, yes: Mac mini M5 Pro 48 GB, Mac Studio M5 Max 48 GB, MacBook Pro M5 Pro 48 GB and 1 more. macOS lets the GPU use only part of unified memory by default (about 37 GB on the Mac mini M5 Pro (48 GB)), and Llama 3.3 70B needs about 44 GB. Running sudo sysctl iogpu.wired_limit_mb=43008 in Terminal raises the limit on a 48 GB Mac until the next restart. From 64 GB up, every Mac Apple sells new holds it inside the default limit.

Which quantization of Llama 3.3 70B should I use?

Q4_K_M (about 40 GB of weights) is the common 4-bit GGUF choice. Q8_0 (about 74 GB) is closer to full quality and needs a 96 GB Mac at the default GPU limit. Q3_K_M and Q2_K shrink it to about 34 and 24 GB at a quality cost. The index’s reference figure (13 tok/s on an M5 Max) is for its 4-bit build.

Cite this page

LLMCheck. (2026). Can I run Llama 3.3 70B on a Mac? llmcheck.net. https://llmcheck.net/can-i-run/llama-33-70b/ (updated ). CC BY 4.0.

How to cite LLMCheck →Data: CC BY 4.0Markdown version

Related

Figures from the LLMCheck index, data version 2026-11-18. Verdicts and context limits are estimates from each model’s catalog memory figure and config.json.